ERCIM News No.45 - April 2001 [contents]
Ecological Modelling meets Interactive Geovisualization at GMD
by Oleg Chertov, Alexander Komarov, Natalia Andrienko, Gennady Andrienko and Peter Gatalsky
In December 2000, two forestry experts and specialists in ecological modeling, Prof. O. Chertov and Dr. A. Komarov from the European Forest Institute, Finland, visited the Knowledge Discovery Team of the Autonomous Intelligent Systems Institute of the GMD. They discussed the prospects and possibilities for future scientific co-operation and for the development of an effective tool for decision-making in modern forestry based on new computer technology. Two weeks of intensive work lead to impressive results.
The concept of Sustainable Forest Management (SFM) brings together different approaches and different branches of science. The need to integrate our knowledge of forest development with broader ecological ideas is evident in Europe, where forests are used for the maintenance of biodiversity, aesthetic values, forest products and various other things. There are many different tasks related to forest ecology, economics, politics etc. but uniting these should be done with the help of mathematical modeling and information technology. Many tasks are linked to the spatial distribution of forest territories among surrounding settlements, roads, and agricultural land with different land-use. If we thoroughly analyze the Criteria & Indicators of SFM then we find that all tasks concerned with forest productivity can be analyzed by simulation dynamics models. Tasks concerned with spatial distribution of forest areas such as ecological corridors and the establishment of protected or recreation zones are currently analyzed only by experts. The main task is therefore how to unite the simulation models with spatial visualization of results to create a Decision Support System (DSS) for SFM.
A comparison of total ecosystem carbon loss (as the difference between total carbon in the first and second scenarios) and carbon removed from the ecosystems as cut wood (lumber) at the end of the simulation. The top figure shows the percentage of the total carbon loss due to removed wood. The lower figure shows removed carbon and total ecosystem loss (tonnes per stand) after cuttings in the first scenario. The figure shows that ecosystem loss of carbon due to clear cutting is about ten times the amount of used wood carbon. The reaction is altering in various forest stands.
We attempted to integrate longerm forest ecosystem modelling with modern techniques of exploratory data analysis. The purpose of the integration was to build a prototype system for 'decision making' in sustainable forestry at landscape level. The system should support the spatially oriented tasks arising from the Criteria & Indicators (C&I) of Sustainable Forest Management. We suggest that analysis of the landscape mosaic be done on the basis of a combination of single stand simulation models and software for visual exploration of spatialemporal data.
We performed a model test on a small forest plot consisting of various stands with two silvicultural regime scenarios; the first is a clear cut system with some ecological restrictions dampening the effect of wood harvesting; the second scenario is a full protection of the forest for fifty years. A combined, spatially explicit forest simulation model EFIMOD II, and the DESCARTES software system designed to support visual exploration of spatially referenced data, were used in the experiment. The results of the experiment were impressive. The test shows the high potential of the integration of forest ecosystem modelling with the visual exploration of data on maps. The visualization allows direct representation of time series and spatial patterns of forest dynamics in graphical form, facilitating analysis of the dynamic trends and of the effectiveness of various silvicultural regimes.
We note the usefulness of representation of spatially distributed time series on maps and the informative value of the spatial combination of different attributes. The prototype system enables us to explore the diversity of ecosystem reactions in various forest compartments and to test various spatial combinations of strategies and zoning of the forest area.
There is a strong demand for a new user-friendly modeling system integrating various types of forest models with exploratory data visualization for methodologically easier and more expressive decision-making based on a longerm simulation at forest enterprise or landscape level. Currently we are preparing a project proposal incorporating these topics to be submitted for EU funding.
Gennady Andrienko - GMD
Tel: +49 2241 142486